PARP Inhibition in Epithelial Ovarian Cancer: High Hopes Undergo a Reality Check

Issues Regarding Future Trials

The clinical trials conducted with PARP inhibitors in EOC have taught us many lessons, but they raise even more questions. What began as a very rational drug development has evolved into an increasingly complex path to drug approval.

1. Multiple PARP inhibitors appear to be active in EOC in phase I and II trials. Currently, there are no data comparing one PARP inhibitor to another in the clinical arena. In contrast to what was previously thought, however, recent data from cell lines suggest that iniparib does not share the same mechanism of action as olaparib and veliparib.[23] Although all three drugs create γH2AX foci, olaparib and veliparib do so via inhibition of PARP1 and PARP2, while iniparib likely inhibits PARP5 and PARP6. The clinical implications of this distinction are unclear. History has taught us that many targeted agents are not as targeted in practice as they seemed during their initial development and that sometimes the actual targets differ substantially from the theoretical ones.

2. The toxicities associated with PARP inhibitors as monotherapy are generally mild, especially when compared with traditional chemotherapy. This has prompted consideration of their use in the chemoprevention setting. The population of BRCA mutation carriers unaffected by cancer will continue to grow as patients, their families, and healthcare providers become increasingly aware of hereditary cancer syndromes. The Australian Ovarian Cancer Study recently documented germline BRCA1/2 mutations in 13% of the patients, but only 57% of the mutation carriers had pedigrees suggestive of a hereditary syndrome.[24] Indeed, the National Comprehensive Cancer Network guidelines now include the recommendation that a diagnosis of EOC, FTC, or PPC generates a referral for genetic counseling and testing, regardless of a woman's family history (www.nccn.org, version 1.2011).

Given the lack of an effective screening strategy for EOC, chemoprevention would be an attractive alternative to the current reliance on risk-reducing salpingo-oophorectomy in BRCA mutation carriers. However, the amount of toxicity from preventive therapy that a young, otherwise healthy woman in her reproductive years can tolerate differs radically from what can be tolerated by an older woman with life-threatening cancer. Data on the effects of long-term exposure to PARP inhibition, especially with respect to potential induction of other malignancies, will be needed before it can be widely accepted as a chemoprevention strategy.

3. The best timing for PARP inhibitor use has yet to be defined with respect to the frontline, maintenance, or recurrent-disease setting. No formal data are available from the frontline setting yet. If we determine that PARP inhibition increases the likelihood of curing EOC, the frontline setting would be preferred. If, as seems more likely, PARP inhibition becomes another tool to prolong survival of EOC but does not cure it, then reserving use for the recurrent setting when minimizing toxicity is even more of a priority seems reasonable.

4. The best strategy for employing PARP inhibitors has yet to be defined with respect to monotherapy (intermittent or continuous), combination with chemotherapy, and/or combination wisth other targeted agents. In particular, appropriate dosing of both the PARP inhibitor and the other agent(s) in a combination approach needs to be determined.

5. The specific population that benefits from PARP inhibitor therapy has yet to be defined. The group whose cancer responds appears to be much broader than simply those with a germline BRCA1/2 mutation, the original population targeted. Expanding the study population to include high-grade serous tumors was the next consideration, with the goal of capturing the hereditary tumors in women who chose not to undergo genetic testing and the sporadic tumors thought to be most similar to BRCA-deficient ones. However, histology alone does not accurately predict BRCA mutation status. One large study found that only 45% of BRCA-mutated EOC had serous histology, although other studies suggest that 10% to 20% may be the more accurate estimate.[4,24,25] Notably, the NCCN guidelines recommend genetic counseling based on the diagnosis of EOC, FTC, or PPC, with no restriction on serous histology. The next line of thinking was that only platinum-sensitive cancers would respond to PARP inhibition. Although more data are needed regarding response to PARP inhibition in platinum-resistant EOC, platinum sensitivity alone does not seem to predict PARP inhibitor response. In the future, the ability to quickly and reliably classify tumors by the presence or absence of a homologous recombination defect (through either germline or somatic mutations) may predict subsequent response to PARP inhibition.

The issue of defining the population best served by PARP inhibition is only one example of a pressing need to prospectively allocate EOC patients to the most appropriate therapy. The Cancer Genome Atlas data have shown that, at least for serous EOC, no single driver mutation (or even small group of mutations) provides easily druggable targets for individualized therapy.[5] This differs from the situation in gastrointestinal stromal tumors, melanoma, and lung cancer, among other examples.

The need for predictive biomarkers only increases as we gain insight into the biology of EOC. Traditionally, all types of EOC, FTC, and PPC have been grouped together in clinical trials and treated similarly. Molecular studies now demonstrate, however, that there are distinct subtypes, including high-grade serous, low-grade serous, clear cell, and mucinous EOC. Debate is ongoing as to where endometrioid EOC tumors belong, since on the one hand, they cluster with high-grade serous tumors in many profiling studies, but on the other hand they often share a gene mutation (ARID1A) and a precursor lesion (endometriosis) with clear cell cancers.[26-29] As a clinical trial community, we will have to design studies that allow a cancer that is already relatively rare to be classified into even smaller groups based on histology and/or molecular pathways; otherwise, we run the risk of only identifying therapies that are effective in the dominant serous subtype and overlooking therapies that work in smaller subpopulations.[29]

6. Drug development presents challenges in the current health care economy, particularly when it is based on identifying a relatively small population that will benefit. Besides the well-known costs of initial research and development followed by the stepwise clinical trial process, multiple issues from the PARP inhibition story are informative about the overall dose-development environment.

The pathway to registration for a PARP inhibitor with the FDA is complex. Orphan drugs can be considered for registration based on phase II results, given the difficulty of mounting phase III trials in rare conditions and the frequent lack of alternative therapies. Although BRCA-deficient tumors are only a subset of EOC and breast cancer as a whole, they are likely too common to qualify for orphan drug eligibility. In addition, since there are other treatment options that offer potential benefit, a randomized phase III trial would likely be needed for registration.[30,31]

Based on the available phase I and II data, olaparib seems to be the PARP inhibitor most poised for phase III development. It currently is formulated as a 50-mg capsule. The recommended dose is 400 mg twice daily, however, meaning that a patient has to swallow 16 capsules a day. In order to simplify dosing, the manufacturer, AstraZeneca, is reformulating olaparib into higher-dose capsules. The reformulation process, however, may delay clinical trial development significantly.

In addition, if a PARP inhibitor's registration is dependent on identifying patients with germline BRCA mutations, the genetic test becomes a companion diagnostic test that requires FDA approval. Myriad Genetics, which holds the patent on commercial BRCA testing in the US, would have to pursue FDA approval distinct from its current Clinical Laboratory Improvements Amendments certification.[31]

A phase III trial with a PARP inhibitor in recurrent EOC would likely be the next step in development because the response in EOC has been more consistent than in breast cancer. The choice of drug for the standard-of-care arm then becomes relevant. Two randomized phase II trials have utilized PLD as the comparator to either olaparib or veliparib (NCT00628251, NCT01113957). Emerging data suggest that the results from these trials might be difficult to interpret due to the high response rate to PLD seen among BRCA mutation carriers. In a heavily pretreated population, Adams et al saw a 57% response rate to PLD among mutation carriers, compared with a rate of 20% among patients with sporadic EOC.[32] This response was associated with significantly improved progression-free and overall survival. Safra et al also documented marked differences in time to treatment failure and overall survival favoring mutation carriers after PLD treatment.[33] Notably, both groups found that PLD response appeared to be independent of platinum sensitivity. When combined with the well-documented improved sensitivity to platinum among mutation carriers, survival calculations in studies including this population must account for their propensity to survive longer, be exposed to additional chemotherapy regimens, and have longer treatment-free intervals than non–mutation carriers.[32] This underscores the need to have as many EOC, FTC, and PPC patients as possible appropriately classified with respect to germline (and possibly somatic) mutation status, not only so that their families can manage their cancer risk proactively, but also so that clinical trial planners can account for mutation carriers in the study design and interpretation.

Summary

PARP inhibitors have been shown to have activity in EOC, FTC, and PPC in the phase I and II setting. Responses have been seen in both BRCA-deficient and sporadic tumors and do not appear to require platinum sensitivity. Although PARP inhibitors have been well-tolerated as monotherapy, additional study is required to determine the efficacy and toxicity of PARP inhibitors in combination with chemotherapy and other targeted agents.

Without question, better understanding of the molecular events that underlie the development of EOC will improve outcomes. The ability to define a high-risk population has contributed significantly to this effort. PARP inhibition started as a rational drug development process to target germline BRCA-mutated malignancies, but has expanded to include tumors with somatic mutations in BRCA1/2 and, potentially, the entire homologous recombination pathway. The traditional concept of treating EOC as a single entity is giving way to recognition that clinical trials will have to address the remarkable heterogeneity in histology and molecular pathways that these tumors present. Unfortunately, the process of separating tumors by histology and altered pathways is not clear-cut, since tumors of different histology can have the same mutation and vice versa.

The incredible motivation of the community of EOC researchers, clinicians, and patients to find better answers for this disease has resulted in remarkably rapid accrual to the trials reviewed here. Although there are significant obstacles to overcome in the continued evaluation of PARP inhibition in EOC, FTC, and PPC, the unprecedented mobilization of the EOC community should bolster efforts to resolve those challenges.

Financial Disclosure: Dr. Zorn is principal investigator at UPMC for NCT01113957, which is sponsored by Abbott.